Validation of modeling and analysis methodology for train-bridge interaction in terms of displacements by a Python-Abaqus Framework

Resumo

This paper proposes a framework for the dynamic analysis of railway bridges. Railways are an important mode of transport, especially suitable for transporting large loads over long distances, or for transporting passengers. Bridges are auxiliary structures for railways, generally made of reinforced concrete or steel, with different types. Over the decades, models to represent the dynamic effect of trains have evolved, from simple moving loads to complex models of dynamic train-bridge interaction. The dynamic analysis of railway bridges presents a challenge: to achieve accurate results from complex models, while maintaining computational efficiency and low operational cost. Computational modeling strategies include considering the vehicle with a multibody formulation to represent the carbody, bogies and wheelsets, connected through flexible elements that simulate the primary and secondary suspensions, and modeling the bridge following the finite element method. The equations of motion of a three-dimensional vehicle were formulated, describing the physical and geometric characteristics and their degrees of freedom. First, the structure is modeled, and modal analysis is performed, which is calibrated in terms of natural frequency. After checking whether the natural frequency of the structure is adequate, integrate it with the forces coming from the vehicle. Wheel resultant forces are determined in a Python script, based on the properties and characteristics of the vehicle and the spectrum of track irregularities. Two examples of application of the train-bridge interaction analysis methodology will be presented. First, a numerical example is performed. Second, an example using experimental data, of real monitoring of a bridge subject to train action. The proximity of results demonstrates the ability to analyze and represent the phenomenon in terms of maximum displacements. This represents a good capacity to analyze the train-bridge interaction combined with a low operational and computational cost methodology. The proposal of this Framework and its validation in numerical and experimental terms is important for the development of in-depth research in the area, such as, for example, to generate a long dataset to allow the training of an artificial neural network or improvement of analysis methods with considerations of more variables.